collaborators

7 papers

cs.LG2026

Model Collapse Is Not a Bug but a Feature in Machine Unlearning for LLMs

Yan Scholten, Sophie Xhonneux, Leo Schwinn +1

Current unlearning methods for LLMs optimize on the private information they seek to remove by incorporating it into their fine-tuning data. We argue this not only risks reinforcin…

cs.CR2026

LLM-Safety Evaluations Lack Robustness

Tim Beyer, Sophie Xhonneux, Simon Geisler +3

In this paper, we argue that current safety alignment research efforts for large language models are hindered by many intertwined sources of noise, such as small datasets, methodol…

cs.CL2025

A Generative Approach to LLM Harmfulness Mitigation with Red Flag Tokens

David Dobre, Mehrnaz Mofakhami, Sophie Xhonneux +2

Many safety post-training methods for large language models (LLMs) are designed to modify the model's behaviour from producing unsafe answers to issuing refusals. However, such dis…

cs.CL2025

Jailbreak Distillation: Renewable Safety Benchmarking

Jingyu Zhang, Ahmed Elgohary, Xiawei Wang +5

Large language models (LLMs) are rapidly deployed in critical applications, raising urgent needs for robust safety benchmarking. We propose Jailbreak Distillation (JBDistill), a no…

cs.LG2025

Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models

Michael Noukhovitch, Shengyi Huang, Sophie Xhonneux +3

The dominant paradigm for RLHF is online and on-policy RL: synchronously generating from the large language model (LLM) policy, labelling with a reward model, and learning using fe…

cs.LG2025

Soft Prompt Threats: Attacking Safety Alignment and Unlearning in Open-Source LLMs through the Embedding Space

Leo Schwinn, David Dobre, Sophie Xhonneux +2

Current research in adversarial robustness of LLMs focuses on discrete input manipulations in the natural language space, which can be directly transferred to closed-source models.…